33 Neuraminidase

Joe C. Wu, Gregory W. Peet, Simon J. Coutts, Robert J. Eckner, Johanna A. Griffin, and Peter R. Farina. Biochemistry , 1995, 34 (21), pp 7154–7160...
1 downloads 0 Views 2MB Size
7154

Biochemistry 1995, 34, 7154-7160

Non-Sialate Inhibitor of Influenza A/WSN/33 Neuraminidase Joe C. Wu,* Gregory W. Peet, Simon J. Coutts, Robert J. Eckner, Johanna A. Griffin, and Peter R. Farina Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road, P.O. Box 368, Ridgefield, Connecticut 06877 Received October IS, 1994; Revised Manuscript Received February 16, 1995@

ABSTRACT: An N1 strain of influenza A virus neuraminidase (A/WSN/33 NA) was purified and used to screen for inhibitors. As a result, a well-known tuberculostatic, 4’-formylacetanilide thiosemicarbazone (or thiacetazone), was identified. Thiacetazone is a non-sialate compound and inhibits the enzyme in a noncompetitive manner with respect to the substrate sialic acid. Mechanistic studies indicate that the inhibition was due to the competition of thiacetazone with Ca2+, which maintains N1 neuraminidase in an active conformation. The Ki for the inhibition was estimated to be about 4 pM. Equilibrium exchange experiments revealed that when purified A/WSN/33 NA was incubated with 5 pM 45CaC12, 2 mol of 45Ca2f ion was exchanged into each mole of NA tetramer and subsequently displaced from the enzyme upon the introduction of the inhibitor. Inhibition of plaque formation by thiacetazone in an MDCK cell culture that had been infected with the influenza A/WSN/33 virus was demonstrated. Thiacetazone was highly specific for A/WSN/33 neuraminidase, since little effect was noted when it was tested against NAs from the other strains of influenza virus or from bacteria. This compound might represent a group of non-sialate inhibitors of influenza NA that bind to a noncatalytic or an allosteric site on the enzyme.

Influenza virus is one of the most important but poorly controlled human pathogens. Some flu outbreaks in this century have caused serious pandemics and resulted in significant loss of life [for review, see Webster et al. (1993), Smith and Palese (1989), and Potter (1992)l. From a molecular and cellular biology standpoint influenza is one of the best-studied viruses [for review, see Katze and Krug (1990) and Lamb (1989)], yet little progress has been made in combating the disease. Although flu vaccines have been used selectively, particularly for the elderly and high-risk groups, the hypermutability of the virus has been a major obstacle that limits the extensive application of vaccines to the general public. Currently, new vaccines need to be formulated each year on the basis of WHO’S best guess of what antigenic determinants are likely to emerge in the next outbreak. Since scale-up of the vaccines for clinical use requires at least 8 months of time, strain prediction thus must be made far in advance of any epidemic. Inevitably, the vaccine has sometimes been mismatched with the epidemic strain (Beyer et al., 1993). Alternatively, amantadine and its structural analogue rimantadine have also been demonstrated to be effective in inhibiting virus replication [Dolin et al., 1982; Wingfield et al., 1969; Kat0 & Eggers, 1969; for review, see Monto and Arden (1992), Hay (1992), and Helenius (1992)l. It was recently reported that the alteration of the MZ channel conductivity by these compounds was the underlying mechanism of inhibition (Pinto et al., 1992). Unfortunately, the rapid emergence of resistant strains (Hayden et al., 1989) and both GI and CNS adverse effects, perhaps due to unanticipated blocking of endogenous ion channels such as the N-methyl-D-aspartic acid receptor-gated ion channel (Kornhuber et al., 1991; Stoof et al., 1992), limit their effectiveness and therapeutic application.

* Author to whom correspondence should be addressed. @

Abstract published in Advance ACS Abstracts, May 1, 1995. 0006-2960/95/0434-7 154$09.00/0

The influenza neuraminidase (NA) is one of the bestcharacterized viral enzymes at the molecular level (Air et al., 1989; Varghese et al., 1983; Bossart-Whitaker et al., 1993). Its enzymatic function is to catalyze the hydrolysis of terminal N-acetylsialic acid from glycoconjugates on the cell surface. It is generally agreed that viral NA plays a pivotal role in allowing the progeny virus to be released from the surface of infected cells, as well as in preventing selfaggregation of the virus (Griffin & Compans, 1979, 1983; Liu & Air, 1993; Schulman & Palese, 1977). The enzyme may also assist in virus spreading and access to new cells by cleaving sialic acid from the mucins that overlay the epithelial cells of the respiratory tract (Klenk & Rott, 1988; Burnet & Stone, 1974). It is, therefore, plausible that the inhibition of viral NA might retard the mobility of the virus and confine the progeny virions to the infected cell surface. Consequently, the spread of the freshly budded virions might be controlled, and the restricted progeny would be an ideal antigen for the immune response. The early work of Meindl, Palese, and co-workers involving the synthesis of sialic acid analogues led them to identify the first NA transition state inhibitor, 2-deoxy-2,3-dehydroN-acetylneuraminic acid (Meindl et al., 1974; Palese et al., 1974). Recent crystallographic data obtained for influenza NA enabled von Itzstein et al. to design a new generation of transition state inhibitors for this enzyme (von Itzstein et al., 1993). Although such inhibitors, which target the active site of the enzyme, should minimize the emergence of resistant strains, they may also be structurally recognized as substrates or inhibitors by cellular NA or other sialic acid-binding enzymes. Consequently, this might result in cellular toxicity, ineffectiveness, or short lifetime of these compounds in vivo. On the other hand, if a non-sialate inhibitor that targets a non-substrate-binding site of the viral NA can be identified, it might offer a better alternative for treatment. To find a lead to test this strategy, we searched our chemical collection through a high-throughput screening assay. Interestingly, a well-known tuberculostatic, 4’-formylacetanilide thiosemi-

0 1995 American Chemical Society

Non-Sialate Inhibitors of Influenza NA carbazone (thiacetazone) (Colston et al., 1978), was identified, In this paper, we report a mechanistic study showing that thiacetazone targets the influenza A/WSN/33 NA at a site that is not the sialate substrate-binding pocket, but might be close to the calcium-binding domain of the enzyme. Since the calcium-binding site is essential for viral enzyme activity, and since this domain does not resemble the well-known calcium-binding EF-hand structure found in many cellular proteins (Persechini et al., 1989; Weinman, 1991), thiacetazone-type inhibitors might provide better selectivity between viral and cellular enzymes.

MATERIALS AND METHODS Materials. 2-(3-Methoxyphenyl)-N-acetyl-~-neuraminic acid (MPN), salmonella and vibrio cholerae neuraminidases, and dideoxyacetylneuraminic acid were purchased from Sigma Chemical Co. 4-Methylumbelliferyl-a-oxo-N-acetylneuraminic acid was from Boehringer Mannheim. Thiacetazone was from Aldrich Chemical Co. or Sigma. The diazonium salt of 4-amino-2,5-dimethoxy-4'-nitroazobenzene (Fast black K salt) was also purchased from Aldrich. Cell culture media were obtained from Gibco Labs. 45CaC12was from New England NuclearDuPont Co. Bio-Gel P-30 polyacrylamide columns were from Bio-Rad. All other chemicals were reagent grade. Pronase was purchased from Calbiochem. Influenza A N1 virus WSN/33 was kindly provided by Matt Bui and Dr. Ari Helenius of Yale University (New Haven, CT) and by Dr. Doris Bucher of New York Medical College. Influenza A N2 virus ( N W S I Tokyo) was the generous gift of Dr. Yoshihiro Kawaoka, St. Jude Children's Hospital (Memphis, TN). The Ng virus (A/NWS/G70C) and the type B virus (BHIUHG) were kindly supplied by Dr. Gillian Air, University of Alabama (Birmingham, AL). PR-8 virus was generously provided by Dr. Graeme Laver, The John Curtin School of Medical Research (Australia). Purification of Influenza A/WSN/33 Virus. Influenza A/WSN/33 virus used for the preparation of neuraminidase was replicated in MDCK cells. Confluent cell monolayers in R- 1500 roller bottles were usually inoculated with the virus at an infectivity of about 0.001 PFU/cell. Twenty minutes after the inoculation, cultures were treated with 4 pg/mL trypsin and incubated at 37 "C for 42 h. Virus titer was usually found to be 2 x lo8 PFU/mL in the culture fluid at the end of incubation. The fluid was clarified with lowspeed centrifugation and further purified with a sucrose step gradient. The purified virus was free of cellular protein as analyzed by SDS-PAGE and monitored by silver staining. The virus was stored in a storage buffer solution (10 mM Tris-HC1, 20% glycerol, and 1 mM DTT, pH 7.7) at -196 "C. Pur$cation of Neuraminidase. Influenza A/WSN/33 neuraminidase was obtained from the virus grown in MDCK cell culture. The structural stability and enzymatic activity of A/WSN/33 N1 neuraminidase are highly dependent on the presence of Ca2+. The purified enzyme denatured irreversibly in the absence or in the presence of low concentrations of Ca2+. To purify the enzyme with good yield and high specific activity, methodology was developed specifically for A/WSN/33 neuraminidase. The purification procedure was adopted from that published by McKimmBreschkin et al. (1991), with slight modifications for this

Biochemistry, Vol. 34, No. 21, 1995 7155

Table 1: Purification Table for the Neuraminidase from Influenza A/WSN/33 Virus protein specific activity purification yield mass (%) (mg) f&mol/mg/min) (-fold) 0.28 1 100 Concentrated virus 10.3 124 pronase treatment 10.1 0.34 11.7 42 113 airfuged supernatant 0.25 25.8 93 91 Bio-Gel P-30 0.085 particular strain. In a typical preparation of the neuraminidase extramembrane domain, 1 mL of 10 mg/mL A/WSN/ 33 virus particles in a Tris-saline buffer (20 mM Tris-HC1, 4 mM KC1, and 140 mM NaC1, pH 7.8) was treated with 100 pg of pronase so that the ratio of virudpronase = 100/1 (w/w). The resultant suspension was incubated in an Eppendorf thermomixer at 35 "C for 65 min with constant shaking. The digested suspension was immediately centrifuged with an air-driven ultracentrifuge at 32 psi for 15 min. Supernatant containing soluble proteins was separated from the pellet that contained the NA-depleted virus particles and then centrifugally gel-filtered through a Bio-Gel P-30 polyacrylamide column that had been preequilibrated with buffer A (glycerol 25% (v/v), 20 mM K*HP04,50 mM KC1,4 mM MgC12, and 5 mM CaC12, pH 5.85). The gel filtration allowed the pronase and short peptides, which were derived from the degradation of hemagglutinin during treatment, to be removed. The NA enzymatic activity recovered from this purification scheme was typically 95-120% of the original activity associated with the virions. Table 1 summarizes the yield and the enzyme specific activity at each purification step. The purified product was found to be nearly homogeneous as analyzed by SDS-PAGE and monitored by Coomassie Blue staining. The NA heads prepared by this procedure could be stored in liquid N2 for at least 3 months without any significant loss of enzyme activity. Neuraminidase Assay. The neuraminidase activity of virus particles or of the purified tetramer heads was assayed by utilizing an artificial substrate, 2-(3-methoxyphenyl)-Nacetyl-D-neuraminic acid (MPN), as described by Sedmak and Grossberg (1981). In our standard procedure, the enzyme was assayed in 60 p L of buffer B containing 20 mM NaH2P04, 50 mM KC1, 4 mM MgC12, 10 p M CaC12, 0.2 mg/mL BSA, and 0.6 mM MPN, pH 5.85. The calcium was sometimes raised to higher concentrations when specified. The assay was run for 60 min at 32 "C, followed by the addition of 60 p L of buffer B (minus BSA) containing 3 mg/mL Fast black K salt (4-amino-2,5-dimethoxy-4'nitroazobenzene) to convert the hydrolysis product, 3-methoxyphenol, to a diazonium complex as described (Sedmak & Grossberg, 1981). The purified enzyme or the virus used for the assay was usually adjusted to a concentration that gave an OD~~O,,,,,between 0.1 and 0.5, such that Beer's law could be applied. The extinction coefficient of the diazonium adduct determined under these experimental conditions was 19 530 cm-' M-', as calibrated from a titration with 3-methoxyphenol. A neuraminidase assay using the fluorescent substrate 4-methylumbelliferyl-a-~-N-acetylneuramin~c acid (MBN) was conducted as described by Potier et al. (1979). The assay was performed at 32 "C in a cuvette containing 500 p L of buffer B in which the substrate MPN was substituted

7156 Biochemistry, Vol. 34, No. 21, 1995

Wu et al.

120

036 5

100

0.24

4 80

P

3

i

-NH

! 60

n

0.12

0

f

c-s

s

0.00 0.0

N;r 40

0.6

1.2

1.8

2.4

3.0

[Substrate MPN], mM

al

20

4

2 1

100

1

B!

0

100

10'

10'

103

nhlacetazone], pM

FIGURE 1 Suecificitv of thiacetazone inhibition. Neuraminidase activity was issayed In a 60 pL solution containing 5 x unit of the designated NA and 10 pM Ca2+ as described in Materials

and Methods. Curves: 1, influenza A/WSN/33 purified NA tetramer heads; 2, influenza A/WSN/33 virus particle containing the NA holoenzyme; 3, influenza APR-8 virus particle; 4, influenza "WSiTokyo virus; 5, Salmonella typhimurium NA; 6, Vibrio cholerae NA. The inhibitor thiacetazone concentration varied from 0.5 to 130 pM as indicated. The chemical identity of thiacetazone is shown. by 0.2 mh4 MBN. The fluorescence at 450 nm (excited at 345 nm) was followed by a Perkin-Elmer fluorescence spectrophotometer (Model 650-10s). The slope of the fluorescence change at 450 nm vs time was used to calculate enzyme activity. Conditions were adjusted so that the slope was proportional to the quantity of the purified NA or the virus added to the assay mixture. Plaque Assay. Madin-Darby canine kidney (MDCK) cells were acquired from the American Type Culture Collection (ATCC No. CCL 34, MDCK NBL-2) and routinely passaged in Eagle's Minimal Essential medium supplemented with 10% heat-inactivated fetal bovine serum, L-glutamine, and penicillin-streptomycin. In preparation for viral inhibition experiments, 35 mm polystyrene tissue culture plates (Falcon, Becton Dickinson Labs) were seeded with approximately 9.5 x lo5 cells per plate and incubated at 37 "C with 5% COZ. Under these conditions, confluent monolayers were obtained within 2-3 days and were used for viral inhibition experiments within 4-7 days. Influenza A/WSN/33 virus that was used for infection had been passaged in chicken eggs and stored in liquid N2. Prior to inoculation of the MDCK monolayers, concentrated virus stocks were diluted in Dulbecco's phosphate-buffered saline (D-PBS) without calcium chloride or magnesium chloride, but containing 5 mg/mL BSA. The inhibition assays were conducted as described by Hayden et al. (1980). Thiacetazone in culture medium with trace DMSO (